COVER SHEET

Mann, Anthony (2005) Opportunities for Improving the Performance and Reducing the Costs of
Bagasse-Fired Boilers. In Hogarth, DM, Eds. Proceedings International Society of Sugar Cane
Technologists Proceedings of the XXV Congress 25, pages pp. 241-247, Guatemala, Guatemala City.


Accessed from http://eprints.qut.edu.au


Copyright 2005 the authors.
FE5
OPPORTUNITIES FOR IMPROVING THE PERFORMANCE AND REDUCING THE COSTS
OF BAGASSE-FIRED BOILERS
By
A. P. MANN, T. F. DIXON, F. PLAZA and J . A. J OYCE
Sugar Research Institute
a.mann@sri.org.au
KEYWORDS: Boilers, bagasse, CFD, erosion, corrosion
Abstract
The challenges faced by bagasse boiler designers mean that
significant compromise and conservatism are present in the design
of most bagasse fired boilers. This paper describes a number of
recent developments that have allowed factories to capitalise on
some of the conservatism in both new and old boiler designs, to
increase steam output, improve combustion stability and reduce
maintenance costs, through low cost retrofits.
The recently developed swirl spreader combustion system has
overcome some of the limitations of conventional spreader firing
technology. The steam generating capacity of a boiler in Australia
has been successfully increased without any combustion stability
problems and reduced grate deposition.
Computational fluid dynamics (CFD) has been successfully used in
a number of cases to reduce the operating and maintenance costs
associated with erosion of convection bank tubes and erosion and
corrosion of air heater tubes. CFD has also been used to improve
boiler efficiency by correcting gas and air flow non uniformities that
have occurred in a number of boilers. More recently CFD has been
extensively used in the design of a new generation, low capital cost
improved performance bagasse-fired boiler.
Introduction
Many bagasse-fired boilers around the world (especially the older ones) are conservatively
designed and were often intended to operate at low efficiency to reduce the costs of
stockpiling and transporting bagasse. Over time many factories expand their crushing
capability, thereby increasing steam demand.
Factory expansion is in many cases limited by the high capital cost of new steam generating
plant. Consequently, there are significant advantages if the capacity of existing boilers can be
increased or if the capital cost of new boilers can be reduced. The conservative design of
many older boilers makes retrofitting changes to increase capacity a feasible option in many
cases.
Cogeneration is becoming an increasingly important source of income for sugar factories.
Boiler efficiency improvements will maximise income from cogeneration. Boiler operating and
maintenance costs are a large proportion of factory running costs. Some of the major
operating and maintenance issues are erosion and corrosion of boiler tubes, carryover of
unburnt fuel and combustion instability.
This paper summarises some of the work that the Sugar Research Institute has carried out
over the last 10 years to address these issues. The recently developed swirl spreader
combustion system, which reduces the capital cost of new boilers and overcomes some of the
limitations that restrict steam capacity expansion, is discussed. The extensive advances in
CFD modelling of bagasse-fired boilers developed by SRI are also described. These
advances have enabled CFD modelling to reach the stage where it can be routinely used to
design baffles for reducing tube erosion and corrosion, and increasing the efficiency of
bagasse-fired boilers. CFD modelling has also been extensively used in the design of a new
generation reduced capital cost, improved performance bagasse-fired boiler.
Swirl spreader combustion system
SRI has been investigating options for upgrading the performance of bagasse-fired boilers for
several years. Dixon and Plaza (1995b) used the CFD code FURNACE to predict the effect
of excess air, bagasse moisture, fuel firing rate and grate air flow distribution on the heat
fluxes, gas temperatures and unburnt fuel levels in a bagasse-fired boiler. Alternative
conventional spreader and secondary air arrangements were also investigated. Dixon et al.
(1998) used the lumped parameter model BLRDSN to predict how performance parameters,
such as the furnace residence time, would change when a boiler is fired above its Maximum
Continuous Rating (MCR). They found that, for the boiler under investigation (Pleystowe No.
2, Australia), the boiler could operate at over 120% MCR before the furnace residence time
reduced to a level that would give rise to excessive carryover of unburnt fuel. Similar
conclusions have been made about other boilers in the Australian industry. In practice,
however, it has been found that for many boilers, when the bagasse feed rate is increased to
above MCR levels, grate deposition and combustion instability increases. Part of this
deterioration in performance arises from limitations with the design of current conventional
spreaders. The swirl spreader combustion system was developed to address performance
issues that limit the performance of conventional spreaders.
The full swirl spreader combustion system was installed on the Proserpine No.4 boiler
(200 t/h MCR) prior to the 2002 crushing season. This installation was the culmination of
technology development over several years. The development sequence progressed through
stages of (i) concept development on a single full-scale spreader under cold flow conditions in
a test facility at SRI during 1999, (ii) hot testing of two Mark I small-capacity spreader units on
the No.1 boiler (18 t/h MCR) at Farleigh mill during 2000, and (iii) hot testing of an
intermediate size boiler retrofit (Marian No.2 boiler, 113 t/h MCR) utilising three Mark II
spreader units during 2001. The Proserpine No.4 boiler was retrofitted with four swirl
spreaders (Mark III design) that were located in positions 2, 3, 4 and 5. The outer two
conventional spreaders (positions 1 and 6) were not altered. Figure 1 shows a front view of
the new swirl spreaders in positions 2 and 3.
Overall combustion performance of the boiler with the swirl spreaders in operation was
generally very good. Modifications to the scrubber modules in the maintenance season to
reduce pressure drop, and thus recover some performance margin on the ID fan, made it
possible for the boiler to achieve steam loads in excess of 230 t/h (MCR 200 t/h) when
bagasse quality was good (<49% moisture). This output was achieved with only minimal
bagasse flow (5% feeder speed) through the two outer conventional spreaders. This small
flow was necessary to prevent radiant overheating of the bagasse feed chutes to these
convectional spreaders. Even when the bagasse quality was poor, there was only localised
grate deposition (near the front and in the rear corners of the furnace) with most of the grate
remaining free of deposits. This is especially encouraging because with the current set-up on
the Proserpine No. 4 boiler, a large proportion of the hot grate combustion air (between one
third and one half of the original grate air flow) is replaced by cold swirl spreader air. It is
planned that for future installations hot air may be used for the swirl spreaders. Testing
showed that overall furnace performance was relatively unaffected by the settings of the swirl
spreaders. Unburnt fuel measurements showed that the efficiency loss due to unburnt fuel
with the boiler operating at an average steam load of 215 t/h was not much higher than the
unburnt fuel efficiency loss measured by the boiler manufacturer shortly after boiler
commissioning at the much lower steam load of 174 t/h.
The demonstrated ability of the Proserpine No. 4 boiler to generate steam loads in excess of
MCR with four swirl spreaders (while six standard spreaders were required previously) means
that there are significant capital cost savings associated with installing swirl spreaders on new
boilers. By increasing combustion stability with high moisture bagasse and reducing grate
deposition when the boiler is operated at steam loads above the boiler MCR, the swirl
spreader combustion system will also assist with upgrading the capacity of existing boilers.

Fig. 1 Front view of the spreaders in positions 2 and 3 on the Proserpine No. 4 boiler
Advances in CFD modelling of bagasse fired boilers
The FURNACE CFD code
Most of the CFD modelling work carried out by SRI in the boiler field has used the FURNACE
code, which was originally developed for pulverised coal fired boilers by the University of
Sydney (Boyd and Kent, 1986). FURNACE is a three dimensional, structured grid CFD code
that can predict the standard gas and particle flow patterns, combustion and radiation in utility
and industrial boilers firing a range of fuels. The code also has time stepping capability. It
has been modified over a number of years to model the flow and combustion processes in
bagasse fired boilers (Luo, 1993; Woodfield et al., 1998; Mann et al., 2001).
Tube erosion modelling
Erosion of boiler tubes is a significant contributor to the maintenance costs of sugar mills. In
some severe cases, boiler convection banks have required a complete re-tube, a major
capital cost. The problem can be particularly severe in boilers with baffled convection banks.
Tube wear rates are the highest where there are high local gas velocities and high local
particle concentrations. Wear rates also depend strongly on the impact angle.
The FURNACE code was first applied to the problem of convection bank tube erosion in 1994
(Dixon and Plaza, 1995a) and since then the code has been successfully used to redesign a
number of convection banks in Australia (Plaza et al., 1999) and overseas. Figure 2 shows
the predicted gas velocity distribution through a baffled convection bank that experienced
wear at the inlet of the convection bank and near the convection bank outlet baffle. Figure 3
shows the corresponding gas velocity distribution through the bank after modifications to
reduce tube wear rates. The success of the investigations has been noted, for example, by
Muchow and Wallis (2002).

Fig. 2 Side elevation view of the predicted gas velocity (m/s) distribution through a
boiler convection bank experiencing tube wear problems.

Fig. 3 Side elevation view of the predicted gas velocity (m/s) distribution through a boiler
convection bank after baffle modifications to reduce tube wear rates.
Modifications have been made to the FURNACE code to enable it to better predict convection
bank heat transfer performance with different baffle configurations (Mann et al., 2003). This
will allow boiler engineers to make more informed decisions about how to balance the
sometimes competing concerns of reducing tube erosion and improving convection bank heat
transfer. Small modifications, that allow the boiler to carry on for a few more years, can also
be made, after which a re-tube, with major changes to the geometry, is recommended.
Tube erosion can also be a problem at the gas side inlet of air heater tubes. The problem is
particularly severe when the gas and particle flow distribution into the air heater tubes is non
uniform and when the flow into the air heater is not parallel to the longitudinal axes of the
tubes. Tube wear rates have been reduced in several cases using FURNACE to design
baffles that correct the flow non-uniformities and make the flow direction parallel to the
longitudinal axes of the air heater tubes.
Air heater corrosion modelling
Air heater tube corrosion has been a major maintenance cost for many years for many
boilers. Past research (Dixon et al., 2000) has demonstrated that the problem has its origin in
gas and air flow distribution patterns and occurs while a boiler is operating. Measurements of
air heater tube temperatures have shown that wall temperatures close to 60C, that is less
than the flue gas dew point temperature, can occur in air heaters where the mean gas exit
temperature is greater than 260C. Negligible corrosion occurs during the start-up phase,
which has been the traditional explanation for the problem. It has been determined that the
uniformity of the gas flow distribution at the air heater inlet is critical for preventing dew point
condensation inside some air heater tubes. This is coupled with the matching air flow
distribution, which can act to accentuate the gas flow deficiencies. The FURNACE CFD code
has been used to design baffles upstream of the gas and air side inlets of air heaters with
significant reductions in dew point corrosion.
In addition to reducing dew point corrosion, the air heater inlet baffles can make the gas and
air flow distribution through an air heater considerably more uniform, improving heat transfer
performance. Global modelling of air heater performance has shown that the air heater heat
transfer rates can be improved. In one case, where the original gas and air flow distributions
into an air heater were highly non uniform, making the flows more uniform using baffles was
predicted to decrease the gas exit temperature from the air heater by more than 15C. This
will significantly improve boiler efficiency.
Further applications of CFD modelling to bagasse-fired boilers
With increasingly stringent limits on the emissions of particulates and gas phase pollutants
such as NO and CO, the focus of CFD capability development is to predict and reduce
pollutant emissions. Concentrations of CO and NO in the furnace have been measured and
the results analysed so that mathematical models for predicting NO and CO concentrations
can be included in the FURNACE code.
SRI researchers also use FURNACE to investigate the effectiveness of equipment that has
not been installed in Australian bagasse boilers but is operating in many overseas
installations. Two examples are the use of spinning hot secondary air in some furnaces and
the use of duct work that encourages the removal of particulates from the gas flow before
entering the dry or wet scrubbers.
Development of a new design bagasse-fired boiler
SRI, in collaboration with the boiler manufacturer Clyde Babcock Hitachi (CBH) has been
investigating new boiler designs for reduced capital cost and improved combustion
performance. The FURNACE code was used to model the effect of various design changes
on boiler performance. The design modifications included changing the aspect ratio of the
furnace as well as the height, the size and placement of the superheaters and the size and
the shape of the furnace bullnose. The option of installing a new configuration of secondary
air nozzles on the front wall of the furnace to enhance the furnace recirculation pattern (to
increase furnace residence time and therefore improve burnout) was also investigated.
Figure 4 shows a side elevation view of the predicted gas velocity distribution through a
standard design high pressure boiler and Figure 5 shows the corresponding plot with some
front wall secondary air added and swirl spreaders installed. The flow from the front wall
secondary air slots and swirl spreaders causes increased recirculation of furnace gases which
should lead to improved combustion performance and reduced unburnt fuel losses. Furnace
heat transfer was also predicted to increase due to higher temperatures lower down in the
furnace. These, and other modelling results, were used with the design and costing
information supplied by CBH to develop a new generation, low capital cost, improved
performance bagasse fired boiler.


Fig. 4 Side elevation view of the predicted gas velocity (m/s) distribution through a standard
design high pressure boiler.

Fig. 5 Side elevation view of the predicted gas velocity (m/s) distribution through a standard
design high pressure boiler with air entering through slots on the front furnace wall.
Conclusions
The developments summarised in this paper show that there is considerable scope for
upgrading the performance of bagasse-fired boilers. The swirl spreader combustion system
has successfully overcome some of the limitations associated with the conventional
spreaders and enabled the steam generating capacity of a large Australian boiler to be
increased to 115% of MCR. Combustion stability was good and grate deposition was
minimised with swirl spreader firing.
CFD modelling has been used to minimise the operating and maintenance costs of a number
of boilers, by reducing erosion of convection bank tubes and erosion and corrosion of air
heater tubes. Further enhancements to the FURNACE code will allow boilers to operate with
reduced emissions of gas phase pollutants. The code has also been an integral tool in the
development of a new generation, low capital cost, improved performance bagasse-fired
boiler.
Acknowledgements
The work described in this paper was funded by syndicates of Australian sugar factories, the
Sugar Research and Development Corporation, the Australian Greenhouse Office and the
Queensland Department of State Development. The authors would also like to acknowledge
the assistance of the staff of Farleigh, Marian and Proserpine sugar mills with the installation
and testing of the various swirl spreader systems.
References
Boyd, R.K. and Kent, J.H. (1986). Three-dimensional furnace computer modelling, 21st
Symposium (International) on Combustion. The Combustion Institute, 21: 265274.
Dixon, T.F., Joyce, K.N. and Treloar, R. (1998). Increasing boiler capacity by dried
bagasse firing. Proc. Aust. Soc. Sugar Cane Technol., 20: 445-452.
Dixon, T.F. and Plaza, F. (1995a). Prediction of erosion in the convection banks of
boilers. Proc. Aust. Soc. Sugar Cane Technol., 17: 307-316.
Dixon, T.F. and Plaza, F. (1995b). Options for upgrading bagasse furnace performance.
Proc. Aust. Soc. Sugar Cane Technol., 17: 317-324.
Dixon, T.F., Plaza, F. and Mann, A.P. (2000), Prevention of air heater corrosion. Proc.
Aust. Soc. Sugar Cane Technol., 22: 491-497.
Luo, M.C. (1993). Combustion of bagasse in a sugar mill boiler. Ph.D. thesis. Univ. of
Qld., Australia.
Mann, A.P., Pennisi, S.N., Dixon, T.F., Novozhilov, V., Kirkpatrick, M., Kent, J.H.,
Sazonov, V. and Zhang, L. (2001), Modelling of boiler tube erosion. Proc. Aust. Soc. Sugar
Cane Technol., 23: 369-375.
Mann, A.P., Plaza, F. and Dixon, T.F. (2003), Improving the prediction of convection bank
heat transfer. Proc. Aust. Soc. Sugar Cane Technol., 25 (CD-ROM).
Muchow, R.C. and Wallis, E.S. (2002), Fostering innovation for the Australian sugar
industry: SRDCs portfolio performance and prospects. Proc. Aust. Soc. Sugar Cane
Technol., 24: 1-11.
Plaza, F., Dixon, T.F., Dickenson, N.L., Fitzmaurice, A.L. and Owens, M. (1999).
Performance of baffled boilers with redesigned convection banks. Proc. Aust. Soc. Sugar
Cane Technol., 21: 432-437.
Woodfield, P.L., Kent, J.H. and Dixon, T.F. (1998). Computational modelling of a
bagasse-fired furnace effects of fuel moisture. Proc. Aust. Soc. Sugar Cane Technol., 20:
458-464.